Location of structural sections from within a highly distorted complex line drawing

Complex and highly distorted line drawings are produced by subjects attempting the Rey Complex Figure test, a clinical test of neuropsychological assessment. However, the marking scheme conventionally employed can be subjective and unreliable. In this paper, the first stages in automating this scoring system are investigated using a robust technique to locate a reduced set of scoring sections and a knowledge-based system that employs spatial metrics and fuzzy approximation techniques. Testing the technique using clinical data produced encouraging results that support the argument that this is a feasible approach for implementing a fully automated system, and that in its current state, can be immediately applied in a semi-automated system.